Research On Low-Complexity Precoding Of Millimeter Wave Massive MIMO System

With the development of mobile communication,the communication requirements of large capacity,high quality and low delay are increasingly high,but the traditional small-scale antenna system can't meet these requirements.Massive MIMO technology is naturally combined with millimeter-wave communication technology to meet the requirements of large capacity,high quality and low delay.However,with the increase in the number of antennas will also bring some new problems,such as hardware cost,too much energy consumption and so on.Precoding technology plays a very important role in solving the above problems.By processing the transmitted signals,users can be guaranteed to receive high quality signals,reduce the delay of data,and improve the performance of the system.However,there are many devices at the base station and the consumption is high.It is of great practical significance to reduce the hardware complexity and the computational complexity of preprocessing.In this paper,the hybrid precoding technology in Massive MIMO system is studied as follows:First,a PS and a low precision DAC are used simultaneously in the hybrid precoding structure.The effects of quantization of low bit PS and low precision DAC on the spectral efficiency of the system are analyzed,and a discrete alternating minimization hybrid precoding scheme is proposed.The power consumption models of the full connection hybrid precoding structure are established,and the energy efficiency is analyzed.Finally,a compromise design scheme of spectral efficiency and power consumption is provided to optimize spectral efficiency and energy efficiency.Second,a hybrid precoding scheme based on two-timescale is proposed.Divided into different frames according to the timeline,each frame contains a number of time slots.An alternative optimization algorithm is proposed to maximize the rate of Massive MIMO system in downlink with limited transmitting power at the base station.The analog precoding is updated only once in each frame,and the digital precoding is updated once in each time slot to reduce the computational complexity.The update problem of the digital precoding is transformed into a WMMSE problem.Then,the digital precoding is optimized based on the low-dimensional effective channel information.SSCA algorithm is used to construct a convex agent function and update the analog precoding based on channel statistics.Finally,the convergence of the proposed algorithm is analyzed,and the performance advantages of the proposed algorithm are verified by simulation experiments.